PhD Defence • Machine Learning | Natural Language Processing • Less is More: Restricted Representations for Better Interpretability and Generalizability
Please note: This PhD defence will take place online.
Zhiying (Gin) Jiang, PhD candidate
David R. Cheriton School of Computer Science
Supervisor: Professor Jimmy Lin
In this thesis, we aim at improving interpretability and generalizability through restricting representations. We choose to approach interpretability by focusing on attribution analysis to understand which features contribute to prediction on BERT, and to approach generalizability by focusing on effective methods in low-data regime.